I’ve been getting peppered with emails, PM’s, facebook messages, smoke signals, you name it about why I’d use iray and/or why anyone else should. Well, it’s just another digital tool in our ever growing software toolbox. While I was eating supper tonight I prepared a few examples showing what I feel are benefits to iray. Read on for more…
Why have I started using iray so much lately? Well, for me and the projects that I’ve been working on lately it’s been pretty useful and can provide some great results. Yes, there are some restrictions/limitations to deal with but IMHO it’s quite useable for the product-vis and to some degree the architectural-vis communities. I hesitate on the architectural-vis side of things because typically architectural interiors can be quite time consuming to render with path-tracer type applications.
This holds especially true on interiors illuminated by a few small light sources or small windows. You’d have to give iray a lot of fast hardware at a scenario like that in order to achieve decent render times. Architectural exteriors are usually pretty good with path-tracer type applications. You get more light/bounced light in areas in a more efficient/faster manner than on an interior.
Character renders will be difficult with iray since currently there isn’t support for a Sub-Surface-Scattering material. So if you’re a character person, you will not be as pleased with iray. Likewise VFX artists will not find it very helpful as it lacks volumetric support at least at this time as well as the general inability to “cheat” things as you can with other rendering applications.
Then there’s the ever popular topic of hardware. OMG, what a whirlwind of opinions that topic opens up these days. It’s almost as though the rendering application wars have moved over to hardware wars. Simply put, right now to efficiently use iray you need a GPU that supports CUDA. From what I’ve read Nvidia cards are the only cards that support CUDA…so you’ll need to use an Nvidia GPU. Which GPU should you use with iray? The one with the highest number of fast CUDA cores available AND has an amount of on-board memory to hold your scenes, texture maps, and render output size AND comes in at a pricetag you can live with.
FWIW I also have an existing thread on the topic of GPU’s with iray HERE. If you haven’t already read it, be sure to check it out for more info on GPU’s, scene size, etc..
Before I dive into the meat & potatoes of this post let me first say that this IS NOT a mental ray vs. iray post. I do compare things here but it’s not meant to bash one or the other, I’m merely explaining a point. Both mental ray and iray are great rendering solutions when used on the right scenes.
I thought it might be useful to not only discuss some of the upsides to iray, but show them. First up let me provide some hardware details and such. Hardware used is a whitebox (home built workstation) with (1) Quadro 6000 as the primary GPU, (1) Tesla c2070 as the secondary GPU, and finally (1) Tesla c2050 as the third GPU…and a small nuclear power plant powering it all. Yes, I know such a configuration is extreme overkill for the following renders…but I usually can’t render my actual work scenes on anything less than 4GB. Sometimes I get lucky and a work scene will weigh in around 3GB hence the Tesla c2050.
Anyway, all the images below were rendered at 1200 x 1200 pixels using only the GPUs (no CPU) via the iray manger script. I’ve reduced all but one to 800 x 800 pixels for faster loading times. I’ll explain why I didn’t downsize one image later. I’ll share the render times but if you’re using a newer GTX type card keep in mind that you may actually find your render times to be faster than mine since those are typically clocked faster than the Quadro/Tesla cards. Yes, that stinks for me but IMHO if I can’t render my work scenes on a GPU (like those fast GTX cards) then it’s of no use to me (other than some sweet, sweet gaming).
Finally, here we go. Up first is glossy reflections & refractions (12 minute render time):
They look great with iray and render pretty quickly with the right setup. Sure, there are ways to speed up glossy reflections/refractions in mental ray (env.blur shader), but there are limitations to those options and you may lose time in configuration. With iray you just dial down the glossy values and hit the render button. Also of note is the DOF. That’s basically free with iray meaning there’s no render time hit for using it like you would see in mental ray. Of course there are drawbacks to rendering DOF vs. adding it in post but that’s not the point here.
Here’s another glossy example, this time with metal (3 minute render time):
Next I threw in a cutout map with some metal material to see how the cutout function works with iray (3 minute render time):
Self-illuminated materials are WONDERFUL in iray (6 minute render):
HDR Illumination with iray produces nice illumination and shadows that are difficult to achieve with mental ray (2 minute render time).
Just to help show the HDR illumination & shadows a bit better I switched to another HDR (3 minute render time):
Up next is the image that I didn’t scale down. I thought the details would be better viewed at full-res. This is a HDR illuminated scene with a material that has a bump map applied. As you probably know it’s difficult to get a nicely detailed bump effect out of mental ray when using indirect illumination. Indirect illumination and bump map detail isn’t an issue with iray (3 minute render time):
And the last image I have for this particular post is displacement. Yes, it works with iray as well (4 minute render time):